Model comparison
Claude Sonnet 5 vs Kimi K3
Kimi K3 is the stronger model overall, scoring 59.5 to 54.6 on the Noometry Index. Claude Sonnet 5 costs 1.5× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Last verified . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. Claude Sonnet 5 scores higher in 2 categories and Kimi K3 in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Kimi K3 leads 63.0 to 49.1.
- The biggest single-benchmark swing is Surface Evolver Bench: 60% for Claude Sonnet 5 and 95% for Kimi K3.
- Claude Sonnet 5 is cheaper at $2 / $10 per million input/output tokens, against $3 / $15 for Kimi K3.
- Kimi K3 accepts more context: 1.05M tokens versus 1M.
- Kimi K3 has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 5 | Kimi K3 | |
|---|---|---|
| Provider | Anthropic | Moonshot AI |
| Noometry Index | 54.6 | 59.5 |
| Released | 2026-06-29 | 2026-07-16 |
| Weights | Proprietary | Open |
| Context window | 1M | 1.05M |
| Max output | 128K | 1.05M |
| Input $ / M tokens | $2 | $3 |
| Output $ / M tokens | $10 | $15 |
| Results tracked | 51 | 53 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Kimi K3 leads
Claude Sonnet 5: 55.5 (#26), Kimi K3: 61.0 (#10)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| DeepSWE | 53.8% | 68.5% |
| FrontierCode | 42.7% | 44.2% |
| LMArena WebDev | 1541 | 1654 |
| SciCode | 54.3% | 59.5% |
| WeirdML | 68.8% | 82.6% |
| LMArena Coding | 1483 | 1508 |
| ALE-Bench | 1,463 | 1,524 |
| CursorBench | 34.1% | — |
| FrontierSWE | — | 25.9% |
| GSO | 37.3% | — |
Agentic & Tool Use Too close to call
Claude Sonnet 5: 42.8 (#18), Kimi K3: 41.8 (#20)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| APEX-Agents | 54.5% | 50.6% |
| GBAEval | 65.3% | 48.3% |
| Vending-Bench 2 | 6,378 | 5,165 |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 32% |
| GDP.pdf | — | 19% |
| LMArena Search | 1194 | — |
Reasoning Kimi K3 leads
Claude Sonnet 5: 49.1 (#39), Kimi K3: 63.0 (#17)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| SimpleBench | 60.6% | 60.7% |
| NYT Connections (extended) | 75.1% | 93.6% |
| CritPt | 16.9% | 23.4% |
| Chess Puzzles | 35% | 39% |
| LMArena Hard Prompts | 1461 | 1496 |
| Mystery Game Puzzles | 35% | 26% |
| DTBench | 92.5% | 91.2% |
| LMCA | 50% | 52.7% |
| Surface Evolver Bench | 60% | 95% |
| Epoch Capabilities Index | 156.21 | 157.45 |
| ForecastBench | 61.1 | 61.1 |
| ARC-AGI-2 | — | 60.4% |
| ARC-AGI-1 | — | 94.5% |
| Bench to the Future 3 | 0.14 | — |
Math Kimi K3 leads
Claude Sonnet 5: 66.2 (#27), Kimi K3: 74.2 (#16)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 65.6% | 72.2% |
| FrontierMath Tier 4 | 29.3% | 39% |
| OTIS Mock AIME 2024-2025 | 94.7% | 97.2% |
| ProofBench | 77% | 87% |
| LMArena Math | 1467 | 1491 |
| MathArena Final-Answer Competitions | — | 87.8% |
Knowledge Kimi K3 leads
Claude Sonnet 5: 55.6 (#47), Kimi K3: 63.2 (#21)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| GPQA Diamond | 90.5% | 93.1% |
| SimpleQA Verified | 33.7% | 50.6% |
| LMArena Expert | 1490 | 1521 |
Multimodal Claude Sonnet 5 leads
Claude Sonnet 5: 42.4 (#31), Kimi K3: 37.8 (#70)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| Blueprint-Bench 2 | 24.9% | 29.5% |
| LMArena Vision | 1274 | — |
| Furniture Assembly | — | 34.2% |
| LMArena Document | 1466 | — |
Multilingual Kimi K3 leads
Claude Sonnet 5: 53.8 (#55), Kimi K3: 56.3 (#21)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| LMArena Non-English | 1431 | 1466 |
| LMArena Chinese | 1477 | 1529 |
| LMArena French | 1460 | 1491 |
| LMArena German | 1440 | 1488 |
| LMArena Japanese | 1422 | 1487 |
| LMArena Korean | 1411 | 1458 |
| LMArena Russian | 1451 | 1482 |
| LMArena Spanish | 1437 | 1472 |
Instruction Following Kimi K3 leads
Claude Sonnet 5: 76.3 (#41), Kimi K3: 77.7 (#14)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| LMArena Instruction Following | 1452 | 1483 |
Long Context Kimi K3 leads
Claude Sonnet 5: 44.8 (#55), Kimi K3: 45.8 (#29)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| LMArena Longer Query | 1463 | 1494 |
Writing & Preference Kimi K3 leads
Claude Sonnet 5: 69.2 (#25), Kimi K3: 76.6 (#4)
| Benchmark | Claude Sonnet 5 | Kimi K3 |
|---|---|---|
| LMArena Text | 1442 | 1476 |
| LMArena Creative Writing | 1416 | 1454 |
| EQ-Bench Creative Writing | 1794 | 2082 |
| EQ-Bench 4 | 1236 | 1339 |
| LMArena Multi-Turn | 1454 | 1488 |
Frequently asked questions
Is Claude Sonnet 5 better than Kimi K3?
Kimi K3 is the stronger model overall, scoring 59.5 to 54.6 on the Noometry Index. Claude Sonnet 5 costs 1.5× less per token, which makes it the better buy when Kimi K3's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 5 or Kimi K3?
Claude Sonnet 5 is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; Kimi K3 lists at $3 and $15.
Is Claude Sonnet 5 or Kimi K3 better for coding?
Kimi K3 scores higher on coding benchmarks: 61.0 versus 55.5 in the Noometry coding category.
Which has the bigger context window?
Kimi K3 does, with 1.05M tokens against 1M.
How many benchmarks do Claude Sonnet 5 and Kimi K3 share?
45 benchmarks have published results for both models. Claude Sonnet 5 has 51 scored results on Noometry and Kimi K3 has 53.